Geometric Feature Recognition Research Articles

Utilizing pretrained convolutional neural networks for crack detection and geometric feature recognition in concrete surface images

Accurate detection of concrete cracks and identification of crack geometric characteristics are essential for ensuring structural safety. To address this, a concrete surface crack detection method was developed using pretrained convolutional neural networks (CNNs). Additionally, a comprehensive framework was proposed that integrates the pretrained CNN as a feature extractor with various regression algorithms to recognize concrete crack features, including the crack area, maximum width, and average width. Results demonstrate that the detection accuracy and training speed of modified CNN models based on pretrained networks, such as VGG16 and MobileNet, outperform those of CNN models trained from scratch. Moreover, the established CNNs achieve high accuracy in handling diverse images affected by environmental disturbances and noise. The developed comprehensive framework successfully recognized crack geometric features in concrete surface images. Evaluation of four regression algorithms revealed that support vector regression (SVR) achieved R2 values of 0.863 for predicting crack area and 0.764 for predicting the crack average width, while the XGBoost regression algorithm yielded an R2 of 0.782 for predicting the crack maximum width.

An investigation of geometric feature recognition in 3D ship data

The intelligent recognition of ship geometric features is a prerequisite for enabling computers to automatically generate and deform ship hull surfaces according to requirements, thereby replacing the work of human designers to improve design efficiency. This paper aims to research the recognition of geometric features in three-dimensional ship data using PointNet. To achieve this goal, we first construct two ship point cloud datasets suitable for global feature classification and feature part segmentation of three-dimensional hulls. Subsequently, we conducted recognition capability testing to determine the optimal hyperparameters for identifying ship feature networks. Finally, we employ ship models with non-standard positions to implement data augmentation, enhancing the network's robustness in recognizing the initial positions of ships and achieving rapid cognition of three-dimensional ship geometric features. The findings of this research will provide technical support for ship design based on artificial intelligence technology.

Part Deviation Correction Method Based on Geometric Feature Recognition

Part Deviation Correction Method Based on Geometric Feature Recognition

Enhancing three-dimensional convolutional neural network-based geometric feature recognition for adaptive additive manufacturing: a signed distance field data approach

Abstract In the context of additive manufacturing, the adjustment of process data to individual geometric features offers the potential to further increase manufacturing speed and quality, while being widely underestimated in recent research. Unfortunately, the current non-uniform data handling in the CAD-CAM-Link results in a downstream data loss, that prevents the availability of geometric knowledge from being present at any time to apply the more advanced approaches of adaptive slicing and tool path generation. Automatic detection of various geometric entities would be beneficial for classifying partial surfaces and volumetric ranges to gain customized informational insights of geometric parameterization. In this work, an enhanced approach of geometric deep learning for the analysis of voxelized engineering parts will be presented to align the inference representations to modeling paradigms for complex design models like architected materials. Although the baseline voxel representation offers distinct advantages in detection accuracy, it comes with an adversely large memory footprint. The geometry discretization leads to high resolutions needed to capture various detail levels that prevent the analysis of fine-grained objects. To achieve efficient usage of three-dimensional (3D) deep learning techniques, we propose a 3D-convolutional neural network-based feature recognition approach using signed distance field data to limit the needed resolution. These implicit geometric data leverage the advantages of volumetric convolution while alleviating their disadvantages through the use of the continuous signed distance function. When analyzing computer-aided design data for geometric primitive features, a common application task in surface reconstruction of reverse engineering the proposed methodology, achieves a detection accuracy that is in line with the accuracy values achieved by comparable algorithms. This enables the recognition of fine-grained surface instances. The unambiguous shape information extracted could be used in subsequent adaptive slicing algorithms to achieve individual geometry-based hatch generation.

Target recognition using rotating ultrasonic sensor for an amphibious ROV

Recognition and tracking of nearby obstacle targets in the inshore environment are necessary for the driving of amphibious remotely operated vehicles (ROV). In this study, the underwater target recognition and tracking in front of the ROV was conducted with a cost-effective waterproof rotating ultrasonic sensing system. An empirical detection model was developed to establish the detection characteristics of the ultrasonic sensor in the turbid underwater environment. The feature extraction of the underwater targets was performed with the density-based spatial clustering of applications with noise. The maximum relative error of targets geometric feature recognition is 3.85%. The target tracking algorithm was designed based on the Kalman filter. The rotating ultrasonic sensing system was further applied and validated on a full size amphibious tracked ROV in the turbid underwater environment. The experimental results show that the maximum recognition error during the straight movement is less than 5% after Kalman filter processing. The validity of the DBSCAN and Kalman filtering algorithms is verified during the driving of the amphibious ROV.

Dielectric and geometric feature extraction and recognition method of coal and gangue based on VMD-SVM

Aiming at the problems of difficult feature extraction and low recognition accuracy of coal and gangue, this paper proposes a recognition method based on dielectric and geometric features, which uses variational mode decomposition (VMD) to denoise the response signal to obtain its dielectric features; uses the marked watershed method to segment the image boundary to obtain its geometric features. The dielectric and geometric characteristic parameters of coal and gangue are extracted to construct the classification model of support vector machines (SVM). It takes only 0.001483 s for the secondary SVM classification model to process 60 groups of test samples, and the recognition accuracy is 98.33%, which is superior to the constructed SVM classification model, k-nearest neighbor (KNN) classification model, and back propagation neural network (BP) classification model. This method opens up a new way to solve the problem of coal and gangue identification.

One spatial geometric characteristic identification method of a coal mine working face based on ground movement and deformation monitoring data

ABSTRACT Aiming at the bottleneck problem that plagues illegal mining supervision and the historical investigation of closed/abandoned mine mining, that is, the lack of effective large-scale, fast, accurate, and convenient underground mining space geometric feature recognition technology, this paper proposes a spatial geometric characteristic identification method of a coal mine working face based on ground movement and deformation monitoring data. Combing with the mining subsidence prediction model (probability integral method), the identification method with the working face geometric characteristic parameter B as the parameter and the space movement vectors as the observation values is proposed. And the quantum genetic algorithm for solving the parameters of the observation equation is introduced. The high precision, strong ability of anti-random error, and anti-gross error of the method are proved by the simulation experiments. Finally, the method is used to calculate the characteristic parameters of a real goaf. The obtained results are as follows: m = 3.376 m, α = 4.069°, ϑ = 205.527°, H = 530.844 m, D 3 = 620.844 m, D 1 = 196.375 m, X 0 = 56557.125 m, Y 0 = 31516.063 m, and the fitting error is 96.165 mm. The inversion results can accurately describe the spatial geometric characteristics of the goaf.

A Rule-Based System to Promote Design for Manufacturing and Assembly in the Development of Welded Structure: Method and Tool Proposition

Welding is a consolidated technology used to manufacture/assemble large products and structures. Currently, welding design issues are tackled downstream of the 3D modeling, lacking concurrent development of design and manufacturing engineering activities. This study aims to define a method to formalize welding knowledge that can be reused as a base for the development of an engineering design platform, applying design for assembly method to assure product manufacturability and welding operations (design for welding (DFW)). The method of ontology (rule-based system) is used to translate tacit knowledge into explicit knowledge, while geometrical feature recognition with parametric modeling is adopted to couple geometrical information with the identification of welding issues. Results show how, within the design phase, manufacturing issues related to the welding operations can be identified and fixed. Two metal structures (a jack adapter of a heavy-duty prop and a lateral frame of a bracket structure) fabricated with arc welding processes were used as case studies and the following benefits were highlighted: (i) anticipation of welding issues related to the product geometry and (ii) reduction of effort and time required for the design review. In conclusion, this research moves forward toward the direction of concurrent engineering, closing the gap between design and manufacturing.

Robotized assembly and inspection of composite fuselage panels: the LABOR project approach

The aerospace manufacturing is looking at ways to change their production processes in order to improve costs, flexibility and efficiency. The LABOR project, acronym for Lean robotized AssemBly and cOntrol of composite aeRostructures, aims at introducing robotic solutions for the assembly line of fuselage panels adopting medium size robots equipped with smart tools, Human Robot collaboration approach and a distributed software architecture. The system consists of a jig that holds the panel to be assembled, two 6-axis robots mounted on moving platforms in order to reach the whole panel and real time measurements to perform the quality control of the assembled components. The cell automatically performs the referencing of the robot working area on the basis of the recognition of geometric features of the parts to be coupled (edges, holes, etc.) through the use of 3D smart inspection tools. After the ”one shot” drilling and countersinking operations, the hole is processed to guarantee a high standard of the process. Installation and sealing of the fastener complete the working cycle. Furthermore, an advanced multimodal perception system monitors the collaborative workspace in real time for safe human-robot collaboration (HRC) tasks. The project started in March 2018 as part of the European Clean Sky 2 research program. Three partners – LOCCIONI, UNICAMPANIA and UNISA – are developing the prototype cell in collaboration with LEONARDO S.P.A. that is the Topic Manager.

Parameterized representation and solution method of the lightweight 3D model virtual assembly constraint

To solve the problem of the difficulties in assembly constraint representation and solution in lightweight 3D modeling because of the lack of geometric features in the Web environment, a parameterized representation and solution method, lightweight 3D modeling virtual assembly constraint, was proposed in this paper. The assembly constraint element (ACE) was formulated for the representation of assembly constraint information, which includes the geometric feature group, type of assembly constraint and assembly constraint expression. The geometric features of the lightweight 3D model were formulated by parametric equation containing vectors. For different geometric feature of different types of assembly constraints, the corresponding simultaneous parametric equations were solved to obtain assembly constraint expressions, which were stored in the ACE, and the requirement of the appointed assembly constraint was met with the combination of a graphics method and a numerical method. On the basis of the assembly constraint expression in ACE, the assembly constraint solution was achieved by the homogeneous transformation matrix and the DOF reduction algorithm. Finally, geometric feature recognition and online virtual assembly of the lightweight 3D model was accomplished by the method proposed in this paper based on WebGL. Relevant tests were conducted to demonstrate the feasibility and efficiency of this method.

Algorithms for Recognition of Geometric and Parametric Features in prismatic parts using IGES

Algorithms for Recognition of Geometric and Parametric Features in prismatic parts using IGES

Learned 3D shape descriptors for classifying 3D point cloud models

ABSTRACTRecent hardware advances in the field of 3D imaging are democratizing 3D sensing with tremendous scientific and societal implications. Point cloud processing has traditionally required meshing the point clouds output by 3D cameras followed by surface reconstruction, and geometric feature recognition. However, meshing a point cloud is fundamentally an ill-posed problem, and the definition of a good solution is not general. This, in turn, demands new paradigms for processing point cloud information.In this paper, we focus on the task of classifying 3D point clouds captured with commercial 3D cameras, and we integrate supervised machine learning algorithms with three different yet underexplored shape descriptors, namely Light-Field Descriptors, Angular Radial Transform (ART), and Zernike Descriptors (ZD). We evaluate the classification performance of different machine learning algorithms combined with these different shape descriptors on point cloud models obtained from the Google 3D Warehouse, and w...

An automatic recognition and parameter extraction method for structural planes in borehole image

As a breakthrough in borehole imaging technology, digital panoramic borehole camera technology has been widely employed. The high-resolution panoramic borehole images can accurately reproduce the geometric features of structural planes. However, the detection of these features is usually done manually, which is both time-consuming and introduces human errors. To solve this problem, this paper presents a method for the automatic recognition and parameter extraction of borehole geometric features of camera images. In this method, the image's gray and gradient level, and also their projection on the depth axis are used to identify the locations of structural planes. Afterwards, iterative matching is employed by using a template of sinusoidal function to search for structural planes in the identified image blocks. Finally, optimal sine curves are selected as the feature curves of structural planes, and their related parameters are converted into structural plane parameters required for engineering, such as their positions, dip directions, dip angles and fracture widths. The method can automatically identify all of structural planes throughout the whole borehole camera image in a continuous and rapid manner, and obtain the corresponding structural parameters. It has proven highly reliable, accurate and efficient.

Automation of drafting execution by schemes definitions and feature recognition

ABSTRACTDespite of the widespread of annotated 3D virtual models, it is still important to produce engineering drawings for a correct communication among the people of the design and manufacturing chain. Automatic drafting through meta drawing definitions, called Drafting Schemes, is proposed in this paper to save time when drawing becomes a repetitive task. The approach produces drafts with a predetermined content in terms of views arrangement and dimensioning strategy, and guarantees robust results to the variations of the input geometry. In fact, the geometrical entities referenced by the generated dimensions are retrieved through dedicated search algorithms based on recognition of geometric features of the input model.The approach has been implemented and tested on a family of gear motors showing feasibility and good results in terms of drafting time savings.

新設計CADデータへの既存メッシュフィッティング技術の開発

We developed a mesh morphing technique, which can be used to change the shape of existing analytical mesh models, and generated a mesh model for a new shape by morphing the mesh model for the previous one. Therefore, our method is suitable to use for analysis during the early stage of design because it reduces the amount of time required to generate analytical mesh models. However, this technique is impossible to morph a free-form surface, and to change surface type. Accordingly, we developed the fitting morphing technique for solving the problem. This technique is possible to morph based on specification of the STL data as morphed shapes and indication of fitting nodes and movable nodes. We developed a series of techniques that include three distinct features. (1) Geometric feature recognition algorithm based on region growing method. (2) Geometric feature assignment algorithm on the basis of distance and angle of geometric features between original mesh and morphed shapes. (3) Node fitting and moving algorithm by displacement coefficient calculation using Ferguson curve characteristics. We confirmed that this technique can morphed to free-form surface from cylindrical surface by applying to the piston head model. Furthermore, the technique can morphed to maintain the boundary continuity of the shape by applying to the gear model.

Defects’ geometric feature recognition based on infrared image edge detection

Edge detection is an important technology in image segmentation, feature extraction and other digital image processing areas. Boundary contains a wealth of information in the image, so to extract defects’ edges in infrared images effectively enables the identification of defects’ geometric features. This paper analyzed the detection effect of classic edge detection operators, and proposed fuzzy C-means (FCM) clustering-Canny operator algorithm to achieve defects’ edges in the infrared images. Results show that the proposed algorithm has better effect than the classic edge detection operators, which can identify the defects’ geometric feature much more completely and clearly. The defects’ diameters have been calculated based on the image edge detection results.

Research on Optical Feature Recognition Technology of Multiple Micro-Holes

A rapid quality detection system for multiple micro-holes fabricated by electrochemical machining (ECM) is proposed by studying on the optical features of multiple micro-holes. First, the edge of multiple micro-holes is detected exactly with the digital image processing techniques. Second, the multiple micro-holes are recognized through image skeleton extraction and region segmentation algorithm. Then, the least square circle fit arithmetic; the new roundness measurement algorithm and the surface roughness measurement method are synthetically used in the processing of geometric feature recognition of each micro hole. Finally, the practical results prove that the rapid quality detection system is provided with a fast detection speed and high precision for meeting the quality inspection requirement of multiple micro-holes.

Advancing image quantification methods and tools for analysis of nanoparticle electrokinetics

Image processing methods and techniques for high-throughput quantification of dielectrophoretic (DEP) collections onto planar castellated electrode arrays are developed and evaluated. Fluorescence-based dielectrophoretic spectroscopy is an important tool for laboratory investigations of AC electrokinetic properties of nanoparticles. This paper details new, first principle, theoretical and experimental developments of geometric feature recognition techniques that enable quantification of positive dielectrophoretic (pDEP) nanoparticle collections onto castellated arrays. As an alternative to the geometric-based method, novel statistical methods that do not require any information about array features, are also developed using the quantile and standard deviation functions. Data from pDEP collection and release experiments using 200 nm diameter latex nanospheres demonstrates that pDEP quantification using the statistic-based methods yields quantitatively similar results to the geometric-based method. The development of geometric- and statistic-based quantification methods enables high-throughput, supervisor-free image processing tools critical for dielectrophoretic spectroscopy and automated DEP technology development.

Development of Quality Detecting System for Micro Holes by ECM Based on Machine Vision Technology

A rapid quality detection system for micro hole fabricated by electrochemical machining (ECM) is proposed by studying on the optical features of micro hole. The edge of micro hole is detected exactly with the digital image processing techniques, such as, the middle value filter, the threshold value transformation, and so on. The least square circle fit arithmetic and the new roundness measurement algorithm are synthetically used in the processing of geometric feature recognition of micro holes. The practical results prove that the method with exact conclusion and high precision can satisfy the quality inspection requirement of micro holes fabricated by ECM.

Research and Development of Micro-Hole Quality Detecting System Based on Machine Vision

An effective approach of quality inspection for micro hole is proposed by studying on the geometric feature of micro hole edge. The edge of micro hole is detected exactly with the digital image processing techniques, such as, the median filter, the threshold value transformation, and so on. The least square circle fit arithmetic and a new roundness measurement arithmetic are synthetically used in the processing of geometric feature recognition of micro hole. The practical results prove that the method with exact conclusion and high precision can satisfy the quality inspection requirement of micro hole.